Subject: Re: TCP problems resolved
To: Perry Hutchison <>
From: John Hawkinson <jhawk@MIT.EDU>
List: port-sun3
Date: 12/29/1999 10:00:21
In message <>, Perry Hutchison writes:
>> rfc1323  conflicts with  rfc1144  (vj compression)
>If these are known to conflict, shouldn't the implementation
>be smart enough to prevent enabling both at once?

This is a bit tougher since they operate at different layers,
and rfc1323 is a global option across the whole machine and
header compression is per-interface.

>Best in the sense of easiest to accomplish,
>or is 1323 always a win over 1144?

It should be quite the opposite. Header compression is a big win on
low-bandwidth, low-latency (e.g. dialup) links. RFC1323 (large
windows, timestamps) is hardly a win there at all.  Given
the choice, you should pick rfc1144 over rfc1323 for a dialup